Meeting report
Alain Danet
03 février, 2022
1 Introduction
1.1 Documenting community changes
- Many papers report local species richness and temporal trends
- Abundance (Klink et al. 2020)
- Species richness: (Vellend et al. 2013; Dornelas et al. 2014)
- Other aspects of community structure vary through time:
- Composition: (Dornelas et al. 2014; Hillebrand et al. 2018)
- Functional diversity (Barnagaud et al. 2017)
- Evenness (ref)
1.2 Why do we want to assess community changes ?
- Global changes have affect a tremendous effect communities
- Understanding assembly and disassembly rules
- Give a more complex and complete picture of what is going on ?
1.3 Link between different facets of biodiversity
- Species richness:
- Evenness / richness: (+) (Soininen, Passy, and Hillebrand 2012)
- FDiv / richness: (+) (Mouchet et al. 2010)
- FEve / richness: (O) (Mouchet et al. 2010) … to complete
1.4 Research questions
- What are the drivers of community changes ?
- Climate change: more appearance and disappearance in the upper basin
- More species turnover far from the source because of dispersal ability
- Temporal homogeneization at the basin scale?
- Similary in accordance to distance through distance?
- Product:
- Map of temperature and distance to effect on turnover by basin ?
2 Methods
2.1 Data selection
- Unique protocol and unique unit of abundance by site
- One sampling by year and site
- In each site, samplings should be in the median (more or less one month and half, i.e. 1.5 month)
- 5 samplings by site, span 10 years
loc <- filtered_dataset$location %>%
left_join(filtered_dataset$site_quali, by = "siteid") %>%
st_as_sf(coords = c("longitude", "latitude"),
crs = 4326)fig_sites <- paste0(
"Australia absence because span is superior to 10 years of 100s of sites but number of samplings < 10.")world <- ne_countries(scale = "medium", returnclass = "sf")
bb <- st_bbox(loc)
ggplot(data = world) +
geom_sf() +
geom_sf(data = loc, aes(color = protocol), shape = 1) +
theme(legend.position = "bottom") +
labs(title = paste0("Number of sites: ", nrow(loc)))
Figure 2.1: Australia absence because span is superior to 10 years of 100s of sites but number of samplings < 10.
2.2 Biodiversity
2.2.1 Quantity
- Hill numbers
- \(H^0\): Species richness
- \(H^1\): Shannon
- \(H^2\): Simpson
We used also hill numbers with coverage-based correction proposed by (???)
- Evenness:
- Pielou: \(J = H^1 / log(S)\)
2.2.2 Betadiversity
Betadiversity metrics compares the composition of two communities. In our case, we compare the composition of each time step relatively to the baseline, i.e. the first year of sampling.
- Species exchange index: (Hillebrand et al. 2018)
- Presence-absence: \(J = SER_r = \dfrac{S_{imm} + S_{lost}}{S_{tot}}\)
Relative abundance: \(SER_a = \dfrac{\sum_i (p_i - p^{\prime}_i)^2}{\sum_i p_i^2 + \sum_i p^{\prime2}_i - \sum_i p_i p^{\prime}_i}\)
- \(S_{imm}\), \(S_{lost}\), \(S_{tot}\): number of immigrant, lost and total species respectively
- \(i\): species \(i\), \(p\): relative abundance, \(\prime\): second community
\(J\): Jaccard index
Turnover metrics based on presence/absence can be decomposed in appearance/disappearance and nestedness / turnover.
appearance: \(\dfrac{S_{imm}}{S_{tot}}\), disappearance: \(\dfrac{S_{lost}}{S_{tot}}\)
- Turnover: \(J_t = \dfrac{2 * min(S_{lost}, S_{imm})}{S_{common} + (2 * min(S_{lost}, S_{imm})}\)
Nestedness: \(SER_r - J_t\)
3 Environment (Coming soon)
4 Statistical analysis
- Simple lm by site
5 Results
5.1 Direction of trends
ti <- map_dfr(rigal_trends, ~tabyl_df(x = .x, group = "direction"),
.id = "response"
)
ti %>%
filter(direction != "Total") %>%
select(response, direction, percent) %>%
mutate(response = str_replace_all(response, get_var_replacement())) %>%
pivot_wider(names_from = "direction", values_from = "percent") %>%
kable() %>%
kable_styling(bootstrap_options = c("striped", "bordered", "hover"))| response | stable | increase | decrease |
|---|---|---|---|
| Total abundance | 82.7% | 10.0% | 7.3% |
| Log Total Turnover (jaccard) (codyn) abundance | 80.6% | 11.8% | 7.6% |
| Species richness | 81.1% | 11.5% | 5.7% |
| Log species richness | 79.5% | 11.1% | 5.2% |
| Chao species richness | 83.2% | 9.4% | 6.1% |
| Chao Shannon | 80.7% | 9.8% | 5.7% |
| Chao Simpson | 81.7% | 8.8% | 5.8% |
| Chao Evenness | 85.1% | 6.3% | 4.9% |
| Jaccard (binary, similarity) | 80.8% | 2.4% | 13.3% |
| Horn (binary, similarity) | 80.8% | 2.3% | 13.3% |
| Chao (binary, similarity) | 68.0% | 2.7% | 8.2% |
| SER_a (rel abundance) | 81.0% | 3.2% | 13.8% |
| Total Turnover (jaccard) (codyn) | 77.7% | 13.7% | 2.3% |
| Appearance | 73.2% | 11.8% | 2.9% |
| disAppearance | 65.3% | 7.1% | 2.4% |
| Evenness | 82.5% | 7.6% | 6.2% |
| Shannon | 80.0% | 10.2% | 6.1% |
| Simpson | 81.5% | 9.0% | 5.8% |
| Jaccard (binary, dissimilarity) | 78.4% | 13.1% | 2.2% |
| Nestedness (jaccard) | 81.7% | 7.7% | 3.8% |
| Turnover (jaccard) | 57.6% | 6.3% | 1.9% |
ti <- map_dfr(rigal_trends, ~tabyl_df(x = .x, group = "shape_class"),
.id = "response"
)
ti %>%
filter(shape_class != "Total") %>%
select(response, shape_class, percent) %>%
mutate(response = str_replace_all(response, get_var_replacement())) %>%
pivot_wider(names_from = "shape_class", values_from = "percent") %>%
kable() %>%
kable_styling(bootstrap_options = c("striped", "bordered", "hover"))| response | stable_constant | increase_constant | decrease_constant | stable_convex | stable_concave | decrease_decelerated | increase_accelerated | decrease_accelerated | increase_decelerated | NA_constant |
|---|---|---|---|---|---|---|---|---|---|---|
| Total abundance | 72.7% | 9.1% | 6.2% | 6.0% | 4.1% | 1.0% | 0.8% | 0.1% | 0.1% | NA |
| Log Total Turnover (jaccard) (codyn) abundance | 69.5% | 10.4% | 6.9% | 4.8% | 6.3% | 0.3% | 0.2% | 0.4% | 1.2% | 0.0% |
| Species richness | 70.7% | 9.5% | 4.8% | 4.3% | 6.0% | 0.7% | 0.8% | 0.3% | 1.1% | 1.8% |
| Log species richness | 69.3% | 9.3% | 4.4% | 3.4% | 6.9% | 0.5% | 0.7% | 0.3% | 1.1% | 4.3% |
| Chao species richness | 74.1% | 7.8% | 5.3% | 4.6% | 4.5% | 0.7% | 0.9% | 0.1% | 0.7% | 1.3% |
| Chao Shannon | 71.1% | 8.5% | 5.1% | 4.2% | 5.5% | 0.4% | 0.7% | 0.2% | 0.6% | 3.7% |
| Chao Simpson | 72.2% | 7.8% | 5.2% | 4.3% | 5.2% | 0.5% | 0.6% | 0.1% | 0.5% | 3.7% |
| Chao Evenness | 77.5% | 5.5% | 4.6% | 3.3% | 4.2% | 0.3% | 0.4% | 0.1% | 0.4% | 3.7% |
| Jaccard (binary, similarity) | 72.3% | 2.0% | 11.6% | 5.0% | 3.6% | 0.9% | 0.1% | 0.7% | 0.3% | 3.5% |
| Horn (binary, similarity) | 72.5% | 1.9% | 11.6% | 4.4% | 4.0% | 0.8% | 0.1% | 0.9% | 0.3% | 3.6% |
| Chao (binary, similarity) | 58.5% | 1.6% | 5.2% | 5.2% | 4.4% | 1.7% | 0.1% | 1.3% | 1.1% | 21.0% |
| SER_a (rel abundance) | 71.1% | 2.6% | 11.6% | 5.3% | 4.6% | 1.1% | 0.1% | 1.1% | 0.5% | 1.9% |
| Total Turnover (jaccard) (codyn) | 70.5% | 12.2% | 2.0% | 3.5% | 3.7% | 0.2% | 0.9% | 0.1% | 0.5% | 6.3% |
| Appearance | 66.6% | 10.1% | 2.5% | 3.0% | 3.6% | 0.3% | 1.1% | 0.1% | 0.6% | 12.1% |
| disAppearance | 58.8% | 6.1% | 2.0% | 3.6% | 2.8% | 0.4% | 0.6% | 0.1% | 0.4% | 25.1% |
| Evenness | 73.6% | 6.7% | 5.6% | 4.1% | 4.8% | 0.4% | 0.5% | 0.2% | 0.5% | 3.6% |
| Shannon | 70.3% | 8.9% | 5.6% | 4.0% | 5.7% | 0.3% | 0.6% | 0.2% | 0.7% | 3.7% |
| Simpson | 72.2% | 7.8% | 5.1% | 3.7% | 5.6% | 0.3% | 0.5% | 0.3% | 0.8% | 3.6% |
| Jaccard (binary, dissimilarity) | 70.6% | 11.6% | 1.9% | 3.2% | 4.5% | 0.2% | 0.7% | 0.1% | 0.8% | 6.3% |
| Nestedness (jaccard) | 74.6% | 6.2% | 3.3% | 3.6% | 3.6% | 0.4% | 0.8% | 0.1% | 0.6% | 6.8% |
| Turnover (jaccard) | 53.0% | 5.5% | 1.5% | 2.4% | 2.2% | 0.3% | 0.6% | 0.1% | 0.2% | 34.2% |
5.2 Average
slope <- map_dfr(rigal_trends,
~.x %>%
select(siteid, linear_slope),
.id = "response"
)rigal_trends_df <- map2_dfr(
rigal_trends, names(rigal_trends),
~.x %>% mutate(variable = .y)
) %>%
select(variable, siteid, linear_slope) %>%
pivot_wider(names_from = "variable", values_from = "linear_slope")summary_slope <- slope %>%
group_by(response) %>%
summarise(summ = list(enframe(summary_distribution(linear_slope, na.rm = TRUE)))) %>%
unnest(cols = summ) %>%
pivot_wider(names_from = "name", values_from = "value") %>%
select(response, mean, median, sd)
summary_slope %>%
mutate(response = get_var_replacement()[response]) %>%
kable() %>%
kable_styling(bootstrap_options = c("striped", "bordered", "hover")) %>%
scroll_box(width = "100%", height = "1000px")| response | mean | median | sd |
|---|---|---|---|
| Appearance | 0.0043083 | 0.0013017 | 0.0160590 |
| Chao (binary, similarity) | -0.0039328 | 0.0000000 | 0.0218819 |
| Chao evenness | 0.0021943 | 0.0000000 | 0.0869975 |
| Chao species richness | 0.0160926 | 0.0027879 | 0.1864729 |
| Chao shannon | 0.0109086 | 0.0033085 | 0.1104958 |
| Chao simpson | 0.0077291 | 0.0026013 | 0.0994345 |
| Disappearance | 0.0019926 | 0.0000000 | 0.0127173 |
| Evenness | 0.0010409 | 0.0000280 | 0.0185918 |
| SER_a (rel abundance) | -0.0082370 | -0.0035286 | 0.0273119 |
| Horn (binary, similarity) | -0.0052693 | -0.0032226 | 0.0184560 |
| Jaccard (binary, similarity) | -0.0066054 | -0.0041915 | 0.0211723 |
| Jaccard (binary, dissimilarity) | 0.0066054 | 0.0041915 | 0.0211723 |
| Log species richness | 0.5690175 | 0.1305570 | 3.5608475 |
| Log total abundance | 1.2164360 | 0.5582744 | 8.8216646 |
| Nestedness (jaccard) | 0.0029121 | 0.0001510 | 0.0211512 |
| Shannon | 0.0031355 | 0.0008848 | 0.0287062 |
| Simpson | 0.0013521 | 0.0002939 | 0.0142783 |
| Species richness | 0.0233072 | 0.0038939 | 0.1763604 |
| Total turnover (codyn) | 0.0063009 | 0.0038111 | 0.0197541 |
| Total abundance | 2.2395645 | 0.0955316 | 42.9647716 |
| Turnover (jaccard) | 0.0036933 | 0.0000000 | 0.0230755 |
p_jt_sup_ne <- sum(abs(rigal_trends_df$turnover) > abs(rigal_trends_df$nestedness)) /
nrow(rigal_trends_df)- To compare with Dornelas et al. (2014):
- Species richness:
- Dornelas et al. (2014): 1557 measurements of species richness in two consecutive times, 629 (40%) increase, 624 (40%) decrease, and 304 (20%) do not change.
- Our study: 13% increase, 7% decrease, 78% do not change, median: + 131% species by decade (i.e from 2 to 3 species, 10 to 15 species in 10 years)
- Jaccard:
- 10% species replaced per decade (Dornelas et al. 2014)
- 28% of species replaced per decade (Blowes et al. 2019)
- our study: 4% of species per decade
- Turnover / nestedness:
- Blowes et al. (2019): 97% of study shows turnover exceeding nestedness
- our study: 44%
- Jaccard vs dominance-based SER:
- low correlation: \(0.34\)
- Species richness:
rigal_trends_df %>%
ggplot(aes(x = jaccard , y = hillebrand)) +
geom_point() +
# geom_smooth(method = "gam") +
labs(x = "Jaccard trends (similarity, binary)", y = "Hillebrand trends
(similarity, relative abundance)")
rigal_trends_df_loc <- filtered_dataset$location %>%
left_join(rigal_trends_df, by = "siteid") %>%
st_as_sf(coords = c("longitude", "latitude"),
crs = 4326)5.3 Relationships between temporal trends
knitr::include_graphics(here("doc", "fig", "p_cor_slope_tot.png"))
ti <- expand.grid(
resp1 = unique(slope$response),
resp2 = unique(slope$response)
) %>%
filter(resp2 != resp1) %>%
filter(
resp1 %in% c("chao_richness", "species_nb", "log_species_nb",
"total_abundance", "log_total_abundance")) %>%
mutate_all(as.character) %>%
arrange(resp1)
test <- map2(ti$resp1, ti$resp2,
function(x, y) {
bi <- slope %>%
filter(response %in% c(x, y)) %>%
select(siteid, response, linear_slope) %>%
pivot_wider(names_from = "response", values_from = "linear_slope")
return(bi)
}
)
p_trends_trends <- map(test, function(x) {
l <- colnames(x)
x %>%
ggplot(aes(x = !!sym(l[2]), y = !!sym(l[3]))) +
geom_point() +
geom_smooth(method = "loess") +
labs(
x = get_var_replacement()[l[2]],
y = get_var_replacement()[l[3]]
)
})
names(p_trends_trends) <- map_chr(test, ~colnames(.x)[2])plot_grid(
plotlist = p_trends_trends[names(p_trends_trends) %in% "log_species_nb"],
ncol = 3
)
#> `geom_smooth()` using formula 'y ~ x'
#> `geom_smooth()` using formula 'y ~ x'
#> `geom_smooth()` using formula 'y ~ x'
#> `geom_smooth()` using formula 'y ~ x'
#> `geom_smooth()` using formula 'y ~ x'
#> `geom_smooth()` using formula 'y ~ x'
#> `geom_smooth()` using formula 'y ~ x'
#> `geom_smooth()` using formula 'y ~ x'
#> `geom_smooth()` using formula 'y ~ x'
#> `geom_smooth()` using formula 'y ~ x'
#> `geom_smooth()` using formula 'y ~ x'
#> `geom_smooth()` using formula 'y ~ x'
#> `geom_smooth()` using formula 'y ~ x'
#> `geom_smooth()` using formula 'y ~ x'
#> `geom_smooth()` using formula 'y ~ x'
#> `geom_smooth()` using formula 'y ~ x'
#> `geom_smooth()` using formula 'y ~ x'
#> `geom_smooth()` using formula 'y ~ x'
5.4 Spatial trends
5.5 Rarity
Cf other document
5.6 Spatial scale
- Selection on basin data
5.7 Environment
5.7.1 PCA
library(ade4)
library(factoextra)
#> Welcome! Want to learn more? See two factoextra-related books at https://goo.gl/ve3WBa
tar_load(rigal_slp_df)
slope_df <- rigal_slp_dfres.pca <- dudi.pca(select(slope_df, -siteid),
scannf = FALSE, # Hide scree plot
nf = 5 # Number of components kept in the results
)fviz_eig(res.pca)
p_pca <- map(list(c(1,2), c(2, 3), c(1, 3)),
~fviz_pca_var(res.pca,
axes = .x,
col.var = "contrib", # Color by contributions to the PC
gradient.cols = c("#00AFBB", "#E7B800", "#FC4E07"),
repel = TRUE # Avoid text overlapping
) +
theme(legend.position = "none") + labs(title = "")
)
plot_grid(plotlist = p_pca, ncol = 1)
#> Warning: ggrepel: 8 unlabeled data points (too many overlaps). Consider
#> increasing max.overlaps
5.7.1.1 Environment
tar_load(riveratlas_site)
riveratlas_site <- riveratlas_site %>%
mutate(across(starts_with("tmp_"), ~.x / 10))riv <- riveratlas_site %>%
select(all_of(get_river_atlas_significant_var())) %>%
st_drop_geometry()riv %>%
pivot_longer(cols = everything(), names_to = "variable", values_to = "values") %>%
ggplot(aes(x = values)) +
geom_histogram() +
facet_wrap(~variable, scales = "free_x")
#> `stat_bin()` using `bins = 30`. Pick better value with `binwidth`.
pca_riv <- dudi.pca(df = scale(riv), scannf = FALSE, nf = 3)- We see two main axis (as I found in France):
- First axis related to stream size (distance from source, River area)
- Second axis related to Temperature
fviz_eig(pca_riv)
p_pca <- map(list(c(1,2), c(2, 3), c(1, 3)),
~fviz_pca_var(pca_riv,
axes = .x,
col.var = "contrib", # Color by contributions to the PC
gradient.cols = c("#00AFBB", "#E7B800", "#FC4E07"),
repel = TRUE # Avoid text overlapping
) +
theme(legend.position = "none") + labs(title = "")
)
p_pca[[1]]
#> Warning: ggrepel: 5 unlabeled data points (too many overlaps). Consider
#> increasing max.overlaps
p_pca[[2]]
#> Warning: ggrepel: 6 unlabeled data points (too many overlaps). Consider
#> increasing max.overlaps
p_pca[[3]]
#> Warning: ggrepel: 4 unlabeled data points (too many overlaps). Consider
#> increasing max.overlaps
#### Environment - Trends
tar_load(slp_env)
slp_env_for_pca <- slp_env %>%
select(-all_of(c("siteid", "main_bas", "ecoregion", "geometry"))) %>%
rename(get_river_atlas_significant_var())
slp_env_for_pca$geometry <- NULL
pca_slp_env <- dudi.pca(
df = scale(na.omit(slp_env_for_pca)),
scannf = FALSE, nf = 3)p_slp_env_pca <- map(list(c(1,2), c(2, 3), c(1, 3)),
~fviz_pca_var(pca_slp_env,
axes = .x,
col.var = "contrib", # Color by contributions to the PC
gradient.cols = c("#00AFBB", "#E7B800", "#FC4E07"),
repel = TRUE # Avoid text overlapping
) +
theme(legend.position = "none") + labs(title = "")
)
p_slp_env_pca[[1]]
#> Warning: ggrepel: 23 unlabeled data points (too many overlaps). Consider
#> increasing max.overlaps
p_slp_env_pca[[2]]
#> Warning: ggrepel: 19 unlabeled data points (too many overlaps). Consider
#> increasing max.overlaps
p_slp_env_pca[[3]]
#> Warning: ggrepel: 22 unlabeled data points (too many overlaps). Consider
#> increasing max.overlaps
5.7.2 Model
paste0(var_temporal_trends,
" ~ ",
"dist_up_km + tmp_dc_cyr +
(1 + dist_up_km + tmp_dc_cyr | main_bas)"
) %>%
as.formulalibrary(spaMM)
#> Registered S3 methods overwritten by 'registry':
#> method from
#> print.registry_field proxy
#> print.registry_entry proxy
#> spaMM (Rousset & Ferdy, 2014, version 3.9.25) is loaded.
#> Type 'help(spaMM)' for a short introduction,
#> 'news(package='spaMM')' for news,
#> and 'citation('spaMM')' for proper citation.
tar_load(c(trend_env))
names(trend_env) <- var_temporal_trendsti_trends <- map(var_temporal_trends,
~fitme(
as.formula(paste0("scale(", .x, ") ~ dist_up_km + tmp_c_cyr + (1 | main_bas)")),
data = slp_env[,
colnames(slp_env) %in%
c(.x, "siteid", "dist_up_km", "tmp_c_cyr", "main_bas", "ecoregion")
]
)
)
names(ti_trends) <- var_temporal_trendsci <- map_dfr(ti_trends[names(ti_trends) %in% c("log_total_abundance", "log_species_nb", "jaccard",
"appearance", "disappearance", "turnover",
"nestedness", "hillebrand")], function(v) {
x <- confint(v, c("dist_up_km", "tmp_c_cyr"), verbose = FALSE)
tmp <- map_dfr(x, function(x) {
tm <- x$interval
names(tm) <- c("lower", "upper")
return(tm)
}, .id = "parm"
)
return(tmp)
}, .id = "response")
#> Iterative algorithm converges slowly.
#> Iterative algorithm converges slowly.
ci %>%
kable(.,
caption = "Confidence interval (bootstrap)")| response | parm | lower | upper |
|---|---|---|---|
| log_total_abundance | dist_up_km | 0.0001544 | 0.0004751 |
| log_total_abundance | tmp_c_cyr | -0.0124125 | 0.0108759 |
| log_species_nb | dist_up_km | 0.0000083 | 0.0003255 |
| log_species_nb | tmp_c_cyr | 0.0044921 | 0.0297051 |
| jaccard | dist_up_km | 0.0001336 | 0.0004517 |
| jaccard | tmp_c_cyr | -0.0262087 | -0.0012096 |
| hillebrand | dist_up_km | 0.0000493 | 0.0003720 |
| hillebrand | tmp_c_cyr | -0.0283098 | -0.0062380 |
| appearance | dist_up_km | -0.0002093 | 0.0000992 |
| appearance | tmp_c_cyr | -0.0035101 | 0.0237618 |
| disappearance | dist_up_km | -0.0004606 | -0.0001410 |
| disappearance | tmp_c_cyr | -0.0086671 | 0.0101726 |
| nestedness | dist_up_km | -0.0002258 | 0.0000962 |
| nestedness | tmp_c_cyr | -0.0090806 | 0.0150305 |
| turnover | dist_up_km | -0.0003523 | -0.0000345 |
| turnover | tmp_c_cyr | -0.0015705 | 0.0162845 |
from_data_to_predict <- function(
model_list = NULL,
dataset = NULL,
re = NULL
) {
pred <- na.omit(dataset)
for (i in seq_along(names(model_list))) {
tmp <- predict(model_list[[i]],
re.form = as.formula(re),
intervals = "predVar"
)[,1]
names(tmp) <- NULL
pred[[names(model_list)[i]]] <- tmp
}
return(pred)
}xx <- list(
tmp_c_cyr = seq(min(na.omit(slp_env$tmp_c_cyr)), max(na.omit(slp_env$tmp_c_cyr)), length.out = 10),
dist_up_km = seq(min(na.omit(slp_env$dist_up_km)), max(na.omit(slp_env$dist_up_km)), length.out = 10),
main_bas = unique(na.omit(slp_env$main_bas))
)
new_data <- expand.grid(xx) %>%
as_tibble()
pred <- predict(ti_trends[[1]], newdata = new_data, re.form = re, intervals="predVar")
int <- cbind(pred[,1], attr(pred, "intervals")) %>%
as_tibble() %>%
rename(y = V1)xxx <- cbind(int, new_data) %>%
as_tibble
xxx %>%
ggplot(aes(y = y, x = dist_up_km, color = main_bas)) +
geom_point() +
theme(legend.position = "none")re <- paste0("~ tmp_c_cyr + (1 | main_bas)")
pred_tmp <- from_data_to_predict(
model_list = ti_trends,
dataset = slp_env,
re = re
)
#>
|
| | 0%
|
|ssssssssssssssssssssssssss | 41%
|
|ssssssssssssssssssssssssssssssssssssssssssssssssssss | 81%
|
|ssssssssssssssssssssssssssssssssssssssssssssssssssssssssssssssss| 100%
#>
|
| | 0%
|
|ssssssssssssssssssssssssss | 41%
|
|ssssssssssssssssssssssssssssssssssssssssssssssssssss | 81%
|
|ssssssssssssssssssssssssssssssssssssssssssssssssssssssssssssssss| 100%
#>
|
| | 0%
|
|ssssssssssssssssssssssssss | 41%
|
|ssssssssssssssssssssssssssssssssssssssssssssssssssss | 81%
|
|ssssssssssssssssssssssssssssssssssssssssssssssssssssssssssssssss| 100%
#>
|
| | 0%
|
|ssssssssssssssssssssssssss | 41%
|
|ssssssssssssssssssssssssssssssssssssssssssssssssssss | 81%
|
|ssssssssssssssssssssssssssssssssssssssssssssssssssssssssssssssss| 100%
#>
|
| | 0%
|
|ssssssssssssssssssssssssss | 41%
|
|ssssssssssssssssssssssssssssssssssssssssssssssssssss | 81%
|
|ssssssssssssssssssssssssssssssssssssssssssssssssssssssssssssssss| 100%
#>
|
| | 0%
|
|ssssssssssssssssssssssssss | 41%
|
|ssssssssssssssssssssssssssssssssssssssssssssssssssss | 81%
|
|ssssssssssssssssssssssssssssssssssssssssssssssssssssssssssssssss| 100%
#>
|
| | 0%
|
|ssssssssssssssssssssssssss | 41%
|
|ssssssssssssssssssssssssssssssssssssssssssssssssssss | 81%
|
|ssssssssssssssssssssssssssssssssssssssssssssssssssssssssssssssss| 100%
#>
|
| | 0%
|
|ssssssssssssssssssssssssss | 41%
|
|ssssssssssssssssssssssssssssssssssssssssssssssssssss | 81%
|
|ssssssssssssssssssssssssssssssssssssssssssssssssssssssssssssssss| 100%
#>
|
| | 0%
|
|ssssssssssssssssssssssssss | 41%
|
|ssssssssssssssssssssssssssssssssssssssssssssssssssss | 81%
|
|ssssssssssssssssssssssssssssssssssssssssssssssssssssssssssssssss| 100%
#>
|
| | 0%
|
|ssssssssssssssssssssssssss | 41%
|
|ssssssssssssssssssssssssssssssssssssssssssssssssssss | 81%
|
|ssssssssssssssssssssssssssssssssssssssssssssssssssssssssssssssss| 100%
#>
|
| | 0%
|
|ssssssssssssssssssssssssss | 41%
|
|ssssssssssssssssssssssssssssssssssssssssssssssssssss | 81%
|
|ssssssssssssssssssssssssssssssssssssssssssssssssssssssssssssssss| 100%
#>
|
| | 0%
|
|ssssssssssssssssssssssssss | 41%
|
|ssssssssssssssssssssssssssssssssssssssssssssssssssss | 81%
|
|ssssssssssssssssssssssssssssssssssssssssssssssssssssssssssssssss| 100%
#>
|
| | 0%
|
|ssssssssssssssssssssssssss | 41%
|
|ssssssssssssssssssssssssssssssssssssssssssssssssssss | 81%
|
|ssssssssssssssssssssssssssssssssssssssssssssssssssssssssssssssss| 100%
#>
|
| | 0%
|
|ssssssssssssssssssssssssss | 41%
|
|ssssssssssssssssssssssssssssssssssssssssssssssssssss | 81%
|
|ssssssssssssssssssssssssssssssssssssssssssssssssssssssssssssssss| 100%
#>
|
| | 0%
|
|ssssssssssssssssssssssssss | 41%
|
|ssssssssssssssssssssssssssssssssssssssssssssssssssss | 81%
|
|ssssssssssssssssssssssssssssssssssssssssssssssssssssssssssssssss| 100%
#>
|
| | 0%
|
|ssssssssssssssssssssssssss | 41%
|
|ssssssssssssssssssssssssssssssssssssssssssssssssssss | 81%
|
|ssssssssssssssssssssssssssssssssssssssssssssssssssssssssssssssss| 100%
#>
|
| | 0%
|
|ssssssssssssssssssssssssss | 41%
|
|ssssssssssssssssssssssssssssssssssssssssssssssssssss | 81%
|
|ssssssssssssssssssssssssssssssssssssssssssssssssssssssssssssssss| 100%
#>
|
| | 0%
|
|ssssssssssssssssssssssssss | 41%
|
|ssssssssssssssssssssssssssssssssssssssssssssssssssss | 81%
|
|ssssssssssssssssssssssssssssssssssssssssssssssssssssssssssssssss| 100%
#>
|
| | 0%
|
|ssssssssssssssssssssssssss | 41%
|
|ssssssssssssssssssssssssssssssssssssssssssssssssssss | 81%
|
|ssssssssssssssssssssssssssssssssssssssssssssssssssssssssssssssss| 100%
#>
|
| | 0%
|
|ssssssssssssssssssssssssss | 41%
|
|ssssssssssssssssssssssssssssssssssssssssssssssssssss | 81%
|
|ssssssssssssssssssssssssssssssssssssssssssssssssssssssssssssssss| 100%
#>
|
| | 0%
|
|ssssssssssssssssssssssssss | 41%
|
|ssssssssssssssssssssssssssssssssssssssssssssssssssss | 81%
|
|ssssssssssssssssssssssssssssssssssssssssssssssssssssssssssssssss| 100%
p_tmp <- map(
c(
"log_total_abundance", "log_species_nb", "jaccard",
"appearance", "disappearance", "turnover",
"nestedness", "hillebrand"),
function(x) {
pred_tmp %>%
ggplot(aes(
y = !!sym(x),
x = tmp_c_cyr,
color = main_bas
)) +
geom_line() +
theme(legend.position = "none")
}
)
plot_grid(plotlist = p_tmp, ncol = 3)
re <- paste0("~ dist_up_km + (1 | main_bas)")
pred <- na.omit(slp_env)
for (i in seq_along(var_temporal_trends)) {
tmp <- predict(trend_env[[i]],
re.form = as.formula(re)
)[, 1]
names(tmp) <- NULL
pred[[var_temporal_trends[i]]] <- tmp
}
p <- map(
c("log_total_abundance", "log_species_nb", "jaccard",
"appearance", "disappearance", "turnover",
"nestedness", "hillebrand"),
function(x) {
pred %>%
ggplot(aes(
y = !!sym(x),
x = dist_up_km,
color = as.factor(main_bas),
shape = as.factor(ecoregion)
)) +
geom_line() +
theme(legend.position = "none")
}
)
plot_grid(plotlist = p, ncol = 3)
re <- paste0("~ tmp_dc_cyr + (1 | main_bas)")
pred_tmp <- from_data_to_predict(
model_list = trend_env,
dataset = slp_env,
re = re
)
p_tmp <- map(
c("log_species_nb", "jaccard", "appearance", "disappearance", "turnover", "hillebrand"),
function(x) {
pred_tmp %>%
ggplot(aes(
y = !!sym(x),
x = tmp_dc_cyr,
color = as.factor(main_bas)
)) +
geom_line() +
theme(legend.position = "none")
}
)
plot_grid(plotlist = p_tmp, ncol = 3)
- Ecoregion:
re <- paste0("~ dist_up_km + (1 + dist_up_km | ecoregion)")
pred_dist_ecoregion <- from_data_to_predict(
model_list = trend_env,
dataset = slp_env,
re = re
)p_dist_ecoregion <- map(
c("log_species_nb", "jaccard", "appearance", "disappearance", "turnover", "hillebrand"),
function(x) {
pred_dist_ecoregion %>%
ggplot(aes(
y = !!sym(x),
x = dist_up_km,
color = as.factor(ecoregion)
)) +
geom_line() +
theme(legend.position = "none")
}
)
plot_grid(plotlist = p_dist_ecoregion, ncol = 3)
re <- paste0("~ tmp_dc_cyr + (1 + tmp_dc_cyr | ecoregion)")
pred_tmp_ecoregion <- from_data_to_predict(
model_list = trend_env,
dataset = slp_env,
re = re
)p_tmp_ecoregion <- map(
c("log_species_nb", "jaccard", "appearance", "disappearance", "turnover", "hillebrand"),
function(x) {
pred_tmp_ecoregion %>%
ggplot(aes(
y = !!sym(x),
x = tmp_dc_cyr,
color = as.factor(main_bas)
)) +
geom_line() +
theme(legend.position = "none")
}
)
plot_grid(plotlist = p_tmp_ecoregion, ncol = 3)
5.8 Analysis
5.9 Reproducibility
Reproducibility receipt
## datetime
Sys.time()
#> [1] "2022-02-03 09:57:32 CST"
## repository
if(requireNamespace('git2r', quietly = TRUE)) {
git2r::repository()
} else {
c(
system2("git", args = c("log", "--name-status", "-1"), stdout = TRUE),
system2("git", args = c("remote", "-v"), stdout = TRUE)
)
}
#> Local: main /home/alain/Documents/post-these/isu/RivFishTimeBiodiversityFacets
#> Head: [3deb541] 2022-02-02: spamm model basin as factor
## session info
sessionInfo()
#> R version 4.0.5 (2021-03-31)
#> Platform: x86_64-pc-linux-gnu (64-bit)
#> Running under: Debian GNU/Linux 10 (buster)
#>
#> Matrix products: default
#> BLAS: /home/alain/.Renv/versions/4.0.5/lib/R/lib/libRblas.so
#> LAPACK: /home/alain/.Renv/versions/4.0.5/lib/R/lib/libRlapack.so
#>
#> locale:
#> [1] LC_CTYPE=fr_FR.UTF-8 LC_NUMERIC=C
#> [3] LC_TIME=fr_FR.UTF-8 LC_COLLATE=fr_FR.UTF-8
#> [5] LC_MONETARY=fr_FR.UTF-8 LC_MESSAGES=fr_FR.UTF-8
#> [7] LC_PAPER=fr_FR.UTF-8 LC_NAME=C
#> [9] LC_ADDRESS=C LC_TELEPHONE=C
#> [11] LC_MEASUREMENT=fr_FR.UTF-8 LC_IDENTIFICATION=C
#>
#> attached base packages:
#> [1] parallel stats graphics grDevices utils datasets methods
#> [8] base
#>
#> other attached packages:
#> [1] spaMM_3.9.25 cachem_1.0.4
#> [3] memoise_2.0.0 factoextra_1.0.7
#> [5] ade4_1.7-16 inlatools_0.0.1.9001
#> [7] INLA_21.11.22 sp_1.4-5
#> [9] foreach_1.5.1 Matrix_1.3-2
#> [11] future_1.21.0 slider_0.2.2
#> [13] vegan_2.5-7 lattice_0.20-41
#> [15] permute_0.9-5 codyn_2.0.5
#> [17] janitor_2.1.0 viridis_0.5.1
#> [19] viridisLite_0.3.0 cowplot_1.1.1
#> [21] terra_1.4-22 rnaturalearthdata_0.1.0
#> [23] rnaturalearth_0.1.0 sf_1.0-4
#> [25] scales_1.1.1 kableExtra_1.3.1
#> [27] here_1.0.1 lubridate_1.7.9.2
#> [29] magrittr_2.0.1 forcats_0.5.1
#> [31] stringr_1.4.0 dplyr_1.0.4
#> [33] purrr_0.3.4 readr_2.1.1
#> [35] tidyr_1.1.2 tibble_3.1.6
#> [37] ggplot2_3.3.3 tidyverse_1.3.0
#> [39] tarchetypes_0.3.2 targets_0.8.1
#> [41] conflicted_1.1.0 rmarkdown_2.11
#> [43] nvimcom_0.9-123
#>
#> loaded via a namespace (and not attached):
#> [1] readxl_1.3.1 backports_1.2.1 igraph_1.2.6
#> [4] splines_4.0.5 listenv_0.8.0 digest_0.6.27
#> [7] htmltools_0.5.1.1 fansi_0.5.0 cluster_2.1.1
#> [10] tzdb_0.2.0 openxlsx_4.2.3 globals_0.14.0
#> [13] modelr_0.1.8 prettyunits_1.1.1 colorspace_2.0-0
#> [16] rvest_0.3.6 ggrepel_0.9.1 warp_0.2.0
#> [19] haven_2.3.1 xfun_0.28 callr_3.7.0
#> [22] crayon_1.4.2 jsonlite_1.7.2 iterators_1.0.13
#> [25] glue_1.5.1 registry_0.5-1 gtable_0.3.0
#> [28] webshot_0.5.2 car_3.0-10 abind_1.4-5
#> [31] DBI_1.1.1 rstatix_0.7.0 Rcpp_1.0.6
#> [34] progress_1.2.2 units_0.6-7 proxy_0.4-25
#> [37] foreign_0.8-81 httr_1.4.2 wk_0.5.0
#> [40] ellipsis_0.3.2 pkgconfig_2.0.3 farver_2.0.3
#> [43] sass_0.3.1 dbplyr_2.1.0 utf8_1.2.2
#> [46] tidyselect_1.1.1 labeling_0.4.2 rlang_0.4.12
#> [49] munsell_0.5.0 cellranger_1.1.0 tools_4.0.5
#> [52] cli_3.1.0 generics_0.1.0 broom_0.7.4
#> [55] evaluate_0.14 fastmap_1.1.0 yaml_2.2.1
#> [58] processx_3.5.2 knitr_1.36 fs_1.5.1
#> [61] zip_2.1.1 s2_1.0.7 pbapply_1.5-0
#> [64] nlme_3.1-152 slam_0.1-49 ROI_1.0-0
#> [67] xml2_1.3.2 compiler_4.0.5 rstudioapi_0.13
#> [70] curl_4.3.2 e1071_1.7-4 ggsignif_0.6.1
#> [73] reprex_1.0.0 bslib_0.2.4 stringi_1.7.6
#> [76] highr_0.9 ps_1.6.0 rgeos_0.5-5
#> [79] nloptr_1.2.2.2 classInt_0.4-3 vctrs_0.3.8
#> [82] pillar_1.6.4 lifecycle_1.0.1 jquerylib_0.1.3
#> [85] data.table_1.13.6 R6_2.5.1 bookdown_0.24
#> [88] KernSmooth_2.23-18 gridExtra_2.3 rio_0.5.16
#> [91] parallelly_1.23.0 codetools_0.2-18 boot_1.3-27
#> [94] MASS_7.3-53.1 assertthat_0.2.1 rprojroot_2.0.2
#> [97] withr_2.4.3 mgcv_1.8-34 hms_1.1.1
#> [100] grid_4.0.5 minqa_1.2.4 class_7.3-18
#> [103] snakecase_0.11.0 carData_3.0-4 git2r_0.29.0
#> [106] ggpubr_0.4.0 numDeriv_2016.8-1.1References
Barnagaud, Jean-Yves, Pierre Gaüzère, Benjamin Zuckerberg, Karine Princé, and Jens-Christian Svenning. 2017. “Temporal Changes in Bird Functional Diversity Across the United States.” Oecologia 185 (4): 737–48. https://doi.org/10.1007/s00442-017-3967-4.
Blowes, Shane A., Sarah R. Supp, Laura H. Antão, Amanda Bates, Helge Bruelheide, Jonathan M. Chase, Faye Moyes, et al. 2019. “The Geography of Biodiversity Change in Marine and Terrestrial Assemblages.” Science 366 (6463): 339–45. https://doi.org/10.1126/science.aaw1620.
Dornelas, Maria, Nicholas J. Gotelli, Brian McGill, Hideyasu Shimadzu, Faye Moyes, Caya Sievers, and Anne E. Magurran. 2014. “Assemblage Time Series Reveal Biodiversity Change but Not Systematic Loss.” Science 344 (6181): 296–99. https://doi.org/10.1126/science.1248484.
Hillebrand, Helmut, Bernd Blasius, Elizabeth T. Borer, Jonathan M. Chase, John A. Downing, Britas Klemens Eriksson, Christopher T. Filstrup, et al. 2018. “Biodiversity Change Is Uncoupled from Species Richness Trends: Consequences for Conservation and Monitoring.” Journal of Applied Ecology 55 (1): 169–84. https://doi.org/https://doi.org/10.1111/1365-2664.12959.
Klink, Roel van, Diana E. Bowler, Konstantin B. Gongalsky, Ann B. Swengel, Alessandro Gentile, and Jonathan M. Chase. 2020. “Meta-Analysis Reveals Declines in Terrestrial but Increases in Freshwater Insect Abundances.” Science 368 (6489): 417–20. https://doi.org/10.1126/science.aax9931.
Mouchet, Maud A., Sébastien Villéger, Norman W. H. Mason, and David Mouillot. 2010. “Functional Diversity Measures: An Overview of Their Redundancy and Their Ability to Discriminate Community Assembly Rules.” Functional Ecology 24 (4): 867–76. https://doi.org/10.1111/j.1365-2435.2010.01695.x.
Soininen, Janne, Sophia Passy, and Helmut Hillebrand. 2012. “The Relationship Between Species Richness and Evenness: A Meta-Analysis of Studies Across Aquatic Ecosystems.” Oecologia 169 (3): 803–9. https://doi.org/10.1007/s00442-011-2236-1.
Vellend, Mark, Lander Baeten, Isla H. Myers-Smith, Sarah C. Elmendorf, Robin Beauséjour, Carissa D. Brown, Pieter De Frenne, Kris Verheyen, and Sonja Wipf. 2013. “Global Meta-Analysis Reveals No Net Change in Local-Scale Plant Biodiversity over Time.” Proceedings of the National Academy of Sciences 110 (48): 19456–9.